Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 800, Fri Mar 16 17:22:47 2001 UTC revision 1038, Mon Jan 28 00:18:03 2002 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Allen Leung
17    Date: 2002/01/27
18    Tag: leunga-20020127-nowhere
19    Description:
20    
21       Added a target called nowhere in the configuration scripts.
22       Enabling this will build the MLRISC 'nowhere' tool (for translating
23       programs with where-clauses into legal SML code) during installation.
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2002/01/25 21:27:00 EST
28    Tag: Release_110_38
29    Description:
30    
31    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
32    
33    README will be added later.
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2002/01/25
38    Tag: blume-20020125-ffi
39    Description:
40    
41    A large number of tweaks and improvements to ml-nlffi-lib and
42    ml-nlffigen:
43    
44       - ML represenation types have been streamlined
45       - getter and setter functions work with concrete values, not abstract
46         ones where possible
47       - ml-nlffigen command line more flexible (see README file there)
48       - some bugs have been fixed (hopefully)
49    
50    ----------------------------------------------------------------------
51    Name: Lal George
52    Date: 2002/01/24
53    Tag: george-20020124-risc-ra-interface
54    Description:
55    
56       There is a dramatic simplification in the interface to the
57       register allocator for RISC architectures as a result of making
58       parallel copy instructions explicit.
59    
60    ----------------------------------------------------------------------
61    Name: Matthias Blume
62    Date: 2002/01/22
63    Tag: blume-20020122-x86-ccalls
64    Description:
65    
66    Bug fix for c-calls on x86 (having to do with how char- and
67    short-arguments are being handled).
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume
71    Date: 2002/01/21
72    Tag: blume-20020121-ff
73    Description:
74    
75    Another day of fiddling with the FFI...
76    
77    1. Bug fix/workaround:  CKIT does not complain about negative array
78       dimensions, so ml-nlffigen has to guard itself against this possibility.
79       (Otherwise a negative dimension would send it into an infinite loop.)
80    
81    2. Some of the abstract types (light objects, light pointers, most "base"
82       types) in structure C are now eqtypes.
83    
84    3. Added constructors and test functions for NULL function pointers.
85    
86    ----------------------------------------------------------------------
87    Name: Matthias Blume
88    Date: 2002/01/18
89    Tag: blume-20020118-ready-for-new-release
90    Description:
91    
92    Made config/srcarchiveurl point to a new place.  (Will provide boot
93    files shortly.)
94    
95    Maybe we christen this to be 110.38?
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume
99    Date: 2002/01/18
100    Tag: blume-20020118-more-ffifiddle
101    Description:
102    
103    Today's FFI fiddling:
104    
105      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
106        C.Set.* that use concrete (MLRep.*) arguments and results instead
107        of abstract ones.
108    
109      - Provided word-style bit operations etc. for "int" representation
110        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
111        and Long.
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2002/01/18
116    Tag: blume-20020118-use-x86-fp
117    Description:
118    
119    Now that x86-fast-fp seems to be working, I turned it back on again
120    by default.  (Seems to work fine now, even with the FFI.)
121    
122    Other than that, I added some documentation about the FFI to
123    src/ml-nlffigen/README and updated the FFI test examples in
124    src/ml-nlffi-lib/Tests/*.
125    
126    ----------------------------------------------------------------------
127    Name: Allen Leung
128    Date: 2002/01/17
129    Tag: leunga-20020117-x86-fast-fp-call
130    Description:
131    
132       1. Fixed a problem with handling return fp values when x86's fast fp
133          mode is turned on.
134    
135       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
136          of %f32.
137    
138       3. Added a constructor INT32lit to the ast of MLRISC tools.
139    
140    ----------------------------------------------------------------------
141    Name: Matthias Blume
142    Date: 2002/01/16
143    Tag: blume-20020116-ffifiddle
144    Description:
145    
146    More fiddling with the FFI interface:
147    
148     - Make constness 'c instead of rw wherever possible.  This eliminates
149       the need for certain explicit coercions.  (However, due to ML's
150       value polymorphism, there will still be many cases where explicit
151       coercions are necessary.  Phantom types are not the whole answer
152       to modeling a subtyping relationship in ML.)
153    
154     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
155    
156     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
157       clumsy workaround for varargs problem.)
158    
159    ----------------------------------------------------------------------
160    Name: Lal George
161    Date: 2002/01/15
162    Tag: <none>
163    Description:
164    
165    1. Since COPY instructions are no longer native to the architecture,
166       a generic functor can be used to implement the expandCopies function.
167    
168    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
169       TEXT segment.
170    
171    ----------------------------------------------------------------------
172    Name: Matthias Blume
173    Date: 2002/01/15
174    Tag: blume-20020115-ffiupdates
175    Description:
176    
177    1. Fix for bug resulting in single-precision float values being returned
178       incorrectly from FFI calls.
179    
180    2. Small modifications to C FFI API:
181    
182        - memory-allocation routines return straight objects (no options)
183          and raise an exception in out-of-memory situations
184        - unsafe extensions to cast between function pointers and pointers
185          from/to ints
186        - added structure C_Debug as an alternative to structure C where
187          pointer-dereferencing (|*| and |*!) always check for null-pointers
188        - added open_lib' to DynLinkage;  open_lib' works like open_lib
189          but also takes a (possibly empty) list of existing library handles
190          that the current library depends on
191    
192    ----------------------------------------------------------------------
193    Name: Matthias Blume
194    Date: 2002/01/10
195    Tag: blume-20020110-newffigen
196    Description:
197    
198    1. Updates to portable graph code.
199    
200    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
201       more scalable now so that even huge interfaces such as the one
202       for GTK compile in finite time and space. :-)
203       See src/ml-nlffigen/README for details on what's new.
204    
205    ----------------------------------------------------------------------
206    Name: Lal George
207    Date: 2001/01/09 14:31:35 EST 2002
208    Tag: george-20011206-rm-native-copy
209    Description:
210    
211            Removed the native COPY and FCOPY instructions
212            from all the architectures and replaced it with the
213            explicit COPY instruction from the previous commit.
214    
215            It is now possible to simplify many of the optimizations
216            modules that manipulate copies. This has not been
217            done in this change.
218    
219    ----------------------------------------------------------------------
220    Name: Lal George
221    Date: 2001/12/06 16:50:13 EST 2001
222    Tag: george-20011206-mlrisc-instruction
223    Description:
224    
225    Changed the representation of instructions from being fully abstract
226    to being partially concrete. That is to say:
227    
228      from
229            type instruction
230    
231      to
232            type instr                              (* machine instruction *)
233    
234            datatype instruction =
235                LIVE of {regs: C.cellset, spilled: C.cellset}
236              | KILL of {regs: C.cellset, spilled: C.cellset}
237              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
238              | ANNOTATION of {i: instruction, a: Annotations.annotation}
239              | INSTR of instr
240    
241    This makes the handling of certain special instructions that appear on
242    all architectures easier and uniform.
243    
244    LIVE and KILL say that a list of registers are live or killed at the
245    program point where they appear. No spill code is generated when an
246    element of the 'regs' field is spilled, but the register is moved to
247    the 'spilled' (which is present, more for debugging than anything else).
248    
249    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
250    We used to generate:
251    
252            DEFFREG f1
253            f1 := f2 + f3
254            trapb
255    
256    but now generate:
257    
258            f1 := f2 + f3
259            trapb
260            LIVE {regs=[f1,f2,f3], spilled=[]}
261    
262    Furthermore, the DEFFREG (hack) required that all floating point instruction
263    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
264    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
265    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
266    this architecture).
267    
268    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
269    all the architectures. This will result in further simplification of the
270    register allocator that must be aware of them for coalescing purposes, and
271    will also simplify certain aspects of the machine description that provides
272    callbacks related to parallel copies.
273    
274    ANNOTATION should be obvious, and now INSTR represents the honest to God
275    machine instruction set!
276    
277    The <arch>/instructions/<arch>Instr.sml files define certain utility
278    functions for making porting easier -- essentially converting upper case
279    to lower case. All machine instructions (of type instr) are in upper case,
280    and the lower case form generates an MLRISC instruction. For example on
281    the alpha we have:
282    
283      datatype instr =
284         LDA of {r:cell, b:cell, d:operand}
285       | ...
286    
287      val lda : {r:cell, b:cell, d:operand} -> instruction
288        ...
289    
290    where lda is just (INSTR o LDA), etc.
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume
294    Date: 2001/11/22 21:40:00 EST
295    Tag: Release_110_37
296    Description:
297    
298    Release 110.37.  This time for real.
299    
300    ----------------------------------------------------------------------
301    Name: Matthias Blume
302    Date: 2001/11/21 16:35:00 EST
303    Tag: blume-20011121-foot-in-mouth
304    Description:
305    
306    Removed the "Release_110_37" tag because of a serious bug.
307    This will be re-tagged once the bug is fixed.
308    
309    ----------------------------------------------------------------------
310    Name: Matthias Blume
311    Date: 2001/11/21 16:14:00 EST
312    Tag: blume-20011121-forgottenfile
313    Description:
314    
315    Forgot to add a file.  (Just a .tex-file -- part of
316    the CM manual source.)
317    
318    ----------------------------------------------------------------------
319    Name: Matthias Blume
320    Date: 2001/11/21 16:10:00 EST
321    Tag: blume-20011121-invalid_110_37
322    Description:
323    
324    Note: I removed the original tag "Release_110_37" from this commit
325          because we found a serious bug in all non-x86 backends.
326          - Matthias
327    
328    1. Modifications to the SML/NJ code generator and to the runtime system
329       so that code object name strings are directly inserted into code
330       objects at code generation time.  The only business the runtime system
331       has with this is now to read the name strings on occasions.
332       (The encoding of the name string has also changed somewhat.)
333    
334    2. CM now implements a simple "set calculus" for specifying export lists.
335       In particular, it is now possible to refer to the export lists of
336       other libraries/groups/sources and form unions as well as differences.
337       See the latest CM manual for details.
338    
339    3. An separate notion of "proxy" libraries has again be eliminated from
340       CM's model.  (Proxy libraries are now simply a special case of using
341       the export list calculus.)
342    
343    4. Some of the existing libraries now take advantage of the new set
344       calculus.
345       (Notice that not all libraries have been converted because some
346       of the existing .cm-files are supposed to be backward compatible
347       with 110.0.x.)
348    
349    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
350       "General.exnMessage"!  The former relies on a certain hook to be
351       initialized, and that often does not happen in the stand-alone case.)
352    
353    ----------------------------------------------------------------------
354    Name: Lal George
355    Date: 2001/11/21  13:56:18 EST
356    Tag: george-2001121-pseudo-ops
357    Description:
358    
359      Implemented a complete redesign of MLRISC pseudo-ops. Now there
360      ought to never be any question of incompatabilities with
361      pseudo-op syntax expected by host assemblers.
362    
363      For now, only modules supporting GAS syntax are implemented
364      but more should follow, such as MASM, and vendor assembler
365      syntax, e.g. IBM as, Sun as, etc.
366    
367    ----------------------------------------------------------------------
368    Name: Matthias Blume
369    Date: 2001/11/14 11:52:00 EST
370    Tag: blume-20011114-srcname
371    Description:
372    
373    1. Routed the name of the current source file to mlriscgen where it
374       should be directly emitted into the code object.  (This last part
375       is yet to be done.)
376    
377    2. Some cleanup of the pgraph code to make it match the proposal that
378       I put out the other day.  (The proposal notwithstanding, things are
379       still in flux here.)
380    
381    ----------------------------------------------------------------------
382    Name: Lal George
383    Date: 2001/11/14 09:44:04 EST
384    Tag:
385    Description:
386    
387      Fix for a backpatching bug reported by Allen.
388    
389      Because the boundary between short and long span-dependent
390      instructions is +/- 128, there are an astounding number of
391      span-dependent instructions whose size is over estimated.
392    
393      Allen came up with the idea of letting the size of span
394      dependent instructions be non-monotonic, for a maxIter
395      number of times, after which the size must be monotonically
396      increasing.
397    
398      This table shows the number of span-dependent instructions
399      whose size was over-estimated as a function of maxIter, for the
400      file Parse/parse/ml.grm.sml:
401    
402         maxIter            # of instructions:
403            10                      687
404            20                      438
405            30                      198
406            40                        0
407    
408      In compiling the compiler, there is no significant difference in
409      compilation speed between maxIter=10 and maxIter=40. Actually,
410      my measurements showed that maxIter=40 was a tad faster than
411      maxIter=10! Also 96% of the  files in the compiler reach a fix
412      point within 13 iterations, so fixing maxIter at 40, while high,
413      is okay.
414    
415    ----------------------------------------------------------------------
416    Name: Matthias Blume
417    Date: 2001/10/31 15:25:00 EST
418    Tag: blume-20011031-pgraph
419    Description:
420    
421    CKIT:
422    * Changed the "Function" constructor of type Ast.ctype to carry optional
423      argument identifiers.
424    * Changed the return type of TypeUtil.getFunction accordingly.
425    * Type equality ignores the argument names.
426    * TypeUtil.composite tries to preserve argument names but gives up quickly
427      if there is a mismatch.
428    
429    installation script:
430    * attempts to use "curl" if available (unless "wget" is available as well)
431    
432    CM:
433    * has an experimental implementation of "portable graphs" which I will
434      soon propose as an implementation-independent library format
435    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
436    
437    NLFFI-LIB:
438    * some cleanup (all cosmetic)
439    
440    NLFFIGEN:
441    * temporarily disabled the mechanism that suppresses ML output for
442      C definitions whose identifiers start with an underscore character
443    * generate val bindings for enum constants
444    * user can request that only one style (light or heavy) is being used;
445      default is to use both (command-line arguments: -heavy and -light)
446    * fixed bug in handling of function types involving incomplete pointers
447    * generate ML entry points that take record arguments (i.e., using
448      named arguments) for C functions that have a prototype with named
449      arguments
450      (see changes to CKIT)
451    
452    ----------------------------------------------------------------------
453    Name: Allen Leung
454    Date: 2001/10/27 20:34:00 EDT
455    Tag: leunga-20011027-x86-fast-fp-call
456    Description:
457    
458       Fixed the bug described in blume-20010920-slowfp.
459    
460       The fix involves
461          1. generating FCOPYs in FSTP in ia32-svid
462          2. marking a CALL with the appropriate annotation
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume
466    Date: 2001/10/16 11:32:00 EDT
467    Tag: blume-20011016-netbsd
468    Description:
469    
470    Underscore patch from Chris Richards (fixing problem with compiling
471    runtime system under recent NetBSD).
472    
473    ----------------------------------------------------------------------
474    Name: Allen Leung
475    Date: 2001/10/12 17:18:32 EDT 2001
476    Tag: leung-20011012-x86-printflowgraph
477    Description:
478    
479    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
480    
481    ----------------------------------------------------------------------
482    Name: Lal George
483    Date: 2001/10/11 23:51:34 EDT
484    Tag: george-20011011-too-many-instrs
485    Description:
486    
487    The representation of a program point never expected to see more
488    than 65536 instructions in a basic block!
489    
490    ----------------------------------------------------------------------
491    Name: Lal George
492    Date: 2001/10/09 09:41:37 EDT
493    Tag: george-20011008-mlrisc-labels
494    Description:
495    
496    Changed the machine description files to support printing of
497    local and global labels in assembly code, based on host assembler
498    conventions.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2001/09/25 15:25:00 EDT
503    Tag: blume-20010925-exninfo
504    Description:
505    
506    I provided a non-hook implementation of exnName (at the toplevel) and
507    made the "dummy" implementation of exnMessage (at the toplevel) more
508    useful: if nothing gets "hooked in", then at least you are going to
509    see the exception name and a message indicating why you don't see more.
510    
511    [For the time being, programs that need exnMessage and want to use
512    ml-build should either use General.exnMessage (strongly recommended) or
513    refer to structure General at some other point so that CM sees a
514    static dependency.]
515    
516    [Similar remarks go for "print" and "use":  If you want to use their
517    functionality in stand-alone programs generated by ml-build, then use
518    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume
522    Date: 2001/09/20 17:28:00 EDT
523    Tag: blume-20010920-slowfp
524    Description:
525    
526    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume
530    Date: 2001/09/20 17:20:00 EDT
531    Tag: blume-20010920-canonicalpaths
532    Description:
533    
534    0. Updated the BOOT file (something that I forgot to do earlier).
535    
536    1. Small internal change to CM so that it avoids "/../" in filenames
537       as much as possible (but only where it is safe).
538    
539    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
540       that arguments that contain delimiters are passed through correctly.
541       This change also means that all "special" arguments of the form
542       @SMLxxx... must come first.
543    
544    3. Changed install script to put relative anchor names for tool commands
545       into pathconfig.
546    
547    ----------------------------------------------------------------------
548    Name: Matthias Blume
549    >>>>>>> 1.169
550    Date: 2001/09/18 15:35:00 EDT
551    Tag: blume-20010918-readme11036
552    Description:
553    
554    Added README files.
555    
556    ----------------------------------------------------------------------
557    Name: Matthias Blume
558    Date: 2001/09/18 11:45:00 EDT
559    Tag: Release_110_36 (retag)
560    Description:
561    
562    Fixed mistake in config/preloads. Retagged as 110.36.
563    
564    ----------------------------------------------------------------------
565    Name: Matthias Blume
566    Date: 2001/09/18 09:40:00 EDT
567    Tag: Release_110_36_orig (tag changed)
568    Description:
569    
570    New version (110.36).  New bootfiles.
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2001/09/14 16:15:00 EDT
575    Tag: blume-20010914-x86fastfp
576    Description:
577    
578    John committed some changes that Allen made, in particular a (hopefully)
579    correctly working version of the x86-fp module.
580    
581    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
582    flag to "true".  Everything seems to compile to a fixpoint ok, and
583    "mandelbrot" speeds up by about 15%.
584    
585    ----------------------------------------------------------------------
586    Name: Matthias Blume
587    Date: 2001/09/13 11:20:00 EDT
588    Tag: blume-20010913-minimal
589    Description:
590    
591    1. Stefan Monnier's patch to fix a miscompilation problem that
592       was brought to light by John Reppy's work on Moby.
593    
594    2. Implemented a minimal "structure Compiler" that contains just
595       "version" and "architecture".  The minimal version will be
596       available when the full version is not.  This is for backward-
597       compatibility with code that wants to test Compiler.version.
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume
601    Date: 2001/08/28 14:03:00 EDT
602    Tag: blume-20010828-ml-lex
603    Description:
604    
605    Fix for bug 1581, received from Neophytos Michael.
606    
607    ----------------------------------------------------------------------
608    Name: Matthias Blume
609    Date: 2001/08/27 11:20:00 EDT
610    Tag: blume-20010827-readme11035
611    Description:
612    
613    Fleshed out the README file for 110.35.
614    
615    ----------------------------------------------------------------------
616    Name: Matthias Blume
617    Date: 2001/08/24 17:10:00 EDT
618    Tag: Release_110_35
619    Description:
620    
621    New version number (110.35).  New bootfiles.
622    
623    ----------------------------------------------------------------------
624    Name: Lal George
625    Date: 2001/08/24 13:47:18 EDT 2001
626    Tag: george-20010824-MLRISC-graphs
627    Description:
628    
629     removed clusters from MLRISC completely and replaced with graphs.
630    
631    ----------------------------------------------------------------------
632    Name: Matthias Blume
633    Date: 2001/08/23 17:50:00 EDT
634    Tag: blume-20010823-toplevel
635    Description:
636    
637    - some reorganization of the code that implements various kinds of
638      environments in the compiler (static, dynamic, symbolic, combined)
639    - re-implemented the EnvRef module so that evalStream works properly
640      (if the stream contains references to "use", "CM.make", etc.)
641    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume
645    Date: 2001/08/20 15:50 EDT
646    Tag: blume20010820-slipup
647    Description:
648    
649    I forgot to commit a few files.  Here they are...
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume
653    Date: 2001/08/20 15:35:00 EDT
654    Tag: blume-20010820-debugprof
655    Description:
656    
657    !!!! NEW BOOTFILES !!!!
658    
659    This is another round of reorganizing the compiler sources.  This
660    time the main goal was to factor out all the "instrumentation"
661    passes (for profiling and backtracing) into their own library.
662    The difficulty was to do it in such a way that it does not depend
663    on elaborate.cm but only on elabdata.cm.
664    
665    Therefore there have been further changes to both elaborate.cm and
666    elabdata.cm -- more "generic" things have been moved from the former
667    to the latter.  As a result, I was forced to split the assignment
668    of numbers indicating "primtyc"s into two portions: SML-generic and
669    SML/NJ-specific.  Since it would have been awkward to maintain,
670    I bit the bullet and actually _changed_ the mapping between these
671    numbers and primtycs.  The bottom line of this is that you need
672    a new set of bin- and bootfiles.
673    
674    I have built new bootfiles for all architectures, so doing a fresh
675    checkout and config/install.sh should be all you need.
676    
677    The newly created library's name is
678    
679        $smlnj/viscomp/debugprof.cm
680    
681    and its sources live under
682    
683        src/compiler/DebugProf
684    
685    ----------------------------------------------------------------------
686    Name: Matthias Blume
687    Date: 2001/08/15 17:15:00 EDT
688    Tag: blume-20010815-compreorg
689    Description:
690    
691    This is a first cut at reorganizing the CM libraries that make up the
692    core of the compiler.  The idea is to separate out pieces that could
693    be used independently by tools, e.g., the parser, the typechecker, etc.
694    
695    The current status is a step in this direction, but it is not quite
696    satisfactory yet.  Expect more changes in the future.
697    
698    Here is the current (new) organization...
699    
700        What used to be $smlnj/viscomp/core.cm is now divided into
701        six CM libraries:
702    
703             $smlnj/viscomp/basics.cm
704                           /parser.cm
705                           /elabdata.cm
706                           /elaborate.cm
707                           /execute.cm
708                           /core.cm
709    
710        The CM files for these libraries live under src/system/smlnj/viscomp.
711        All these libraries are proxy libraries that contain precisely
712        one CM library component.  Here are the locations of the components
713        (all within the src/compiler tree):
714    
715             Basics/basics.cm
716             Parse/parser.cm
717             ElabData/elabdata.cm
718             Elaborator/elaborate.cm
719             Execution/execute.cm
720             core.cm
721    
722         [This organization is the same that has been used already
723         for a while for the architecture-specific parts of the visible
724         compiler and for the old version of core.cm.]
725    
726         As you will notice, many source files have been moved from their
727         respective original locations to a new home in one of the above
728         subtrees.
729    
730         The division of labor between the new libraries is the following:
731    
732             basics.cm:
733                - Simple, basic definitions that pertain to many (or all) of
734                  the other libraries.
735             parser.cm:
736                - The SML parser, producing output of type Ast.dec.
737                - The type family for Ast is also defined and exported here.
738             elabdata.cm:
739                - The datatypes that describe input and output of the elaborator.
740                  This includes types, absyn, and static environments.
741             elaborator.cm:
742                - The SML/NJ type checker and elaborator.
743                  This maps an Ast.dec (with a given static environment) to
744                  an Absyn.dec (with a new static environment).
745                - This libraries implements certain modules that used to be
746                  structures as functors (to remove dependencies on FLINT).
747             execute.cm:
748                - Everything having to do with executing binary code objects.
749                - Dynamic environments.
750             core.cm:
751                - SML/NJ-specific instantiations of the elaborator and MLRISC.
752                - Top-level modules.
753                - FLINT (this should eventually become its own library)
754    
755    Notes:
756    
757    I am not 100% happy with the way I separated the elaborator (and its
758    data structures) from FLINT.  Two instances of the same problem:
759    
760        1. Data structures contain certain fields that carry FLINT-specific
761           information.  I hacked around this using exn and the property list
762           module from smlnj-lib.  But the fact that there are middle-end
763           specific fields around at all is a bit annoying.
764    
765        2. The elaborator calculates certain FLINT-related information.  I tried
766           to make this as abstract as I could using functorization, but, again,
767           the fact that the elaborator has to perform calculations on behalf
768           of the middle-end at all is not nice.
769    
770        3. Having to used exn and property lists is unfortunate because it
771           weakens type checking.  The other alternative (parameterizing
772           nearly *everything*) is not appealing, though.
773    
774    I removed the "rebinding =" warning hack because due to the new organization
775    it was awkward to maintain it.  As a result, the compiler now issues some of
776    these warnings when compiling init.cmi during bootstrap compilation. On
777    the plus side, you also get a warning when you do, for example:
778       val op = = Int32.+
779    which was not the case up to now.
780    
781    I placed "assign" and "deref" into the _Core structure so that the
782    code that deals with the "lazy" keyword can find them there.  This
783    removes the need for having access to the primitive environment
784    during elaboration.
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blume
788    Date: 2001/08/13
789    Tag: blume-20010813-closures
790    Description:
791    
792    This fix was sent to us by Zhong Shao.  It is supposed to improve the
793    performance of certain loops by avoiding needless closure allocation.
794    
795    ----------------------------------------------------------------------
796    Name: Lal George
797    Date: 2001/07/31 10:03:23 EDT 2001
798    Tag: george-20010731-x86-fmalloc
799    Description: Fixed bug in x86 calls
800    
801        There was a bug where call instructions would mysteriously
802        vanish. The call instruction had to be one that returned
803        a floating point value.
804    
805    ----------------------------------------------------------------------
806    Name: Lal George
807    Date: 2001/07/19 16:36:29 EDT 2001
808    Tag: george-20010719-simple-cells
809    Description:
810    
811    I have dramatically simplified the interface for CELLS in MLRISC.
812    
813    In summary, the cells interface is broken up into three parts:
814    
815      1. CellsBasis : CELLS_BASIS
816    
817            CellsBasis is a top level structure and common for all
818            architectures.  it contains the definitions of basic datatypes
819            and utility  functions over these types.
820    
821      2. functor Cells() : CELLS
822    
823            Cells generates an interface for CELLS that incorporates the
824            specific resources on the target architecture, such as the
825            presence of special register classes, their number and size,
826            and various useful substructures.
827    
828      3. <ARCH>CELLS
829    
830            e.g. SparcCells: SPARCCELLS
831    
832            <ARCH>CELLS usually contains additional bindings for special
833            registers  on the architecture, such as:
834    
835                    val r0 : cell           (* register zero *)
836                    val y : cell            (* Y register *)
837                    val psr : cell          (* processor status register *)
838                    ...
839    
840            The structure returned by applying the Cells functor is opened
841            in this interface.
842    
843    The main implication of all this is that the datatypes for cells is
844    split between CellsBasis and CELLS -- a fairly simple change for user
845    code.
846    
847    In the old scheme the CELLS interface had a definitional binding of
848    the form:
849    
850            signature CELLS = sig
851    
852               structure CellsBasis = CellsBasis
853    
854               ...
855    
856            end
857    
858    With all the sharing constraints that goes on in MLRISC, this old
859    design  quickly leads to errors such as:
860    
861            "structure definition spec inside of sharing ... "
862    
863    
864    and appears to require an unacceptable amount of sharing and where
865    constraint hackery.
866    
867    I think this error message (the interaction of definitional specs and
868    sharing) requires more explanation on our web page.
869    
870    ----------------------------------------------------------------------
871    Name: Matthias Blume
872    Date: 2001/07/19 15:00:00 EDT
873    Tag: blume-20010719-libreorg
874    Description:
875    
876    This update puts together a fairly extensive but straightforward change
877    to the way the libraries that implement the interactive system are
878    organized:
879    
880       The biggest change is the elimination of structure Compiler.  As a
881       replacement for this structure, there is now a CM library
882       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
883       that exports all the substructures of the original structure Compiler
884       directly.  So instead of saying Compiler.Foo.bar one now simply
885       says Foo.bar.  (The CM libraries actually export a collection of
886       structures that is richer than the collection of substructures of
887       structure Compiler.)
888    
889       To make the transition smooth, there is a separate library called
890       $smlnj/compiler/compiler.cm which puts together and exports the
891       original structure Compiler (or at least something very close to it).
892    
893       There are five members of the original structure Compiler
894       that are not exported directly but which instead became members
895       of a new structure Backend (described by signature BACKEND).  These are:
896       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
897       Interact (: INTERACT), structure Machine (: MACHINE), and val
898       architecture (: string).
899    
900       Structure Compiler.Version has become structure CompilerVersion.
901    
902       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
903       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
904       or x86, respectively.
905       Each of these exports the same frontend structures that
906       $smlnj/compiler.cm exports.  But they do not have a structure Backend
907       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
908       Hppa, PPC, Sparc, or X86, respectively.
909    
910       Library $smlnj/compiler/all.cm exports the union of the exports of
911       $smlnj/compiler/<arch>.cm
912    
913       There are no structures <Arch>Compiler anymore, use
914       $smlnj/compiler/<arch>.cm instead.
915    
916       Library host-compiler-0.cm is gone.  Instead, the internal library
917       that instantiates CM is now called cm0.cm.  Selection of the host
918       compiler (backend) is no longer done here but. (Responsibility for it
919       now lies with $smlnj/compiler/current.cm.  This seems to be more
920       logical.)
921    
922       Many individual files have been moved or renamed.  Some files have
923       been split into multiple files, and some "dead" files have been deleted.
924    
925    Aside from these changes to library organization, there are also changes
926    to the way the code itself is organized:
927    
928       Structure Binfile has been re-implemented in such a way that it no
929       longer needs any knowledge of the compiler.  It exclusively deals
930       with the details of binfile layout.  It no longer invokes the
931       compiler (for the purpose of creating new prospective binfile
932       content), and it no longer has any knowledge of how to interpret
933       pickles.
934    
935       Structure Compile (: COMPILE) has been stripped down to the bare
936       essentials of compilation.  It no longer deals with linking/execution.
937       The interface has been cleaned up considerably.
938    
939       Utility routines for dealing with linking and execution have been
940       moved into their own substructures.
941    
942       (The ultimate goal of these changes is to provide a light-weight
943       binfile loader/linker (at least for, e.g., stable libraries) that
944       does not require CM or the compiler to be present.)
945    
946    CM documentation has been updated to reflect the changes to library
947    organization.
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume
951    Date: 2001/07/10 17:30:00 EDT
952    Tag: Release_110_34
953    Description:
954    
955    Minor tweak to 110.34 (re-tagged):
956    
957      - README.html file added to CVS repository
958      - runtime compiles properly under FreeBSD 3.X and 4.X
959    
960    ----------------------------------------------------------------------
961    Name: Matthias Blume
962    Date: 2001/07/10 17:30:00 EDT
963    Tag: Release_110_34
964    Description:
965    
966    New version number (110.34). New bootfiles.
967    
968    ----------------------------------------------------------------------
969    Name: Matthias Blume
970    Date: 2001/07/09 16:00:00 EDT
971    Tag: blume-20010709-more-varargs
972    Description:
973    
974    I changed the handling of varargs in ml-nlffigen again:
975    The ellipsis ... will now simply be ignored (with an accompanying warning).
976    
977    The immediate effect is that you can actually call a varargs function
978    from ML -- but you can't actually supply any arguments beyond the ones
979    specified explicitly.  (For example, you can call printf with its format
980    string, but you cannot pass additional arguments.)
981    
982    This behavior is only marginally more useful than the one before, but
983    it has the advantage that a function or, more importantly, a function
984    type never gets dropped on the floor, thus avoiding follow-up problems with
985    other types that refer to the offending one.
986    
987    ----------------------------------------------------------------------
988    Name: Matthias Blume
989    Date: 2001/07/09 11:25:00 EDT
990    Tag: blume-20010709-varargs
991    Description:
992    
993    1. ckit-lib.cm now exports structure Error
994    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
995       with a warning accompanied by a source location.  Moreover, it
996       merely skips the offending function or type and proceeds with the
997       rest of its work.u  As a result, one can safely feed C code containing
998       "..." to ml-nlffigen.
999    3. There are some internal improvements to CM, providing slightly
1000       more general string substitutions in the tools subsystem.
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume
1004    Date: 2001/06/27 15:10:00 EDT
1005    Tag: blume-20010627-concur
1006    Description:
1007    
1008    Fixed a small bug in CM's handling of parallel compilation.
1009    (You could observe the bug by Control-C-interrupting an ordinary
1010    CMB.make or CM.stabilize and then attaching some compile servers.
1011    The result was that all of a sudden the previously interrupted
1012    compilation would continue on its own.  This was because of
1013    an over-optimization: CM did not bother to clean out certain queues
1014    when no servers were attached "anyway", resulting in the contents
1015    of these queues to grab control when new servers did get attached.)
1016    
1017    There is also another minor update to the CM manual.
1018    
1019    ----------------------------------------------------------------------
1020    Name: Matthias Blume
1021    Date: 2001/06/26 16:15:00 EDT
1022    Tag: blume-20010626-cmdoc
1023    Description:
1024    
1025    Minor typo fixed in CM manual (syntax diagram for libraries).
1026    
1027    ----------------------------------------------------------------------
1028    Name: Matthias Blume
1029    Date: 2001/06/25 22:55:00 EDT
1030    Tag: blume-20010625-x86pc
1031    Description:
1032    
1033    Fixed a nasty bug in the X86 assembly code that caused signal
1034    handlers to fail (crash) randomly.
1035    
1036    ----------------------------------------------------------------------
1037    Name: Matthias Blume
1038    Date: 2001/06/25 12:05:00 EDT
1039    Tag: blume-20010625-nlffigen
1040    Description:
1041    
1042    This update fixes a number of minor bugs in ml-nlffigen as reported by
1043    Nick Carter <nbc@andrew.cmu.edu>.
1044    
1045      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1046      2. Default names for generated files are now derived from the name of
1047         the C file *without its directory*.  In particular, this causes generated
1048         files to be placed locally even if the C file is in some system directory.
1049      3. Default names for generated signatures and structures are also derived
1050         from the C file name without its directory.  This avoids silly things
1051         like "structure GL/GL".
1052         (Other silly names are still possible because ml-nlffigen does not do
1053          a thorough check of whether generated names are legal ML identifiers.
1054          When in doubt, use command line arguments to force particular names.)
1055    
1056    ----------------------------------------------------------------------
1057    Name: Matthias Blume
1058    Date: 2001/06/21 12:25:00 EDT
1059    Tag: blume-20010621-eXene
1060    Description:
1061    
1062    eXene now compiles and (sort of) works again.
1063    
1064    The library name (for version > 110.33) is $/eXene.cm.
1065    
1066    I also added an new example in src/eXene/examples/nbody.  See the
1067    README file there for details.
1068    
1069    ----------------------------------------------------------------------
1070    Name: Matthias Blume
1071    Date: 2001/06/20 16:40:00 EDT
1072    Tag: blume-20010620-cml
1073    Description:
1074    
1075    CML now compiles and works again.
1076    
1077    Libraries (for version > 110.33):
1078    
1079      $cml/cml.cm            Main CML library.
1080      $cml/basis.cm          CML's version of $/basis.cm.
1081      $cml/cml-internal.cm   Internal helper library.
1082      $cml/core-cml.cm       Internal helper library.
1083      $cml-lib/trace-cml.cm  Tracing facility.
1084      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1085    
1086    The installer (config/install.sh) has been taught how to properly
1087    install this stuff.
1088    
1089    ----------------------------------------------------------------------
1090    Name: Matthias Blume
1091    Date: 2001/06/19 17:55:00 EDT
1092    Tag: blume-20010619-instantiate
1093    Description:
1094    
1095    This un-breaks the fix for bug 1432.
1096    (The bug was originally fixed in 110.9 but I broke it again some
1097    time after that.)
1098    
1099    ----------------------------------------------------------------------
1100    Name: Matthias Blume
1101    Date: 2001/06/19 17:25:00 EDT
1102    Tag: blume-20010619-signals
1103    Description:
1104    
1105    This should (hopefully) fix the long-standing signal handling bug.
1106    (The runtime system was constructing a continuation record with an
1107    incorrect descriptor which would cause the GC to drop data on the floor...)
1108    
1109    ----------------------------------------------------------------------
1110    Name: Matthias Blume
1111    Date: 2001/06/15 15:05:00 EDT
1112    Tag: blume-20010615-moresparc
1113    Description:
1114    
1115    Here is a short late-hour update related to Sparc c-calls:
1116    
1117     -- made handling of double-word arguments a bit smarter
1118    
1119     -- instruction selection phase tries to collapse certain clumsily
1120        constructed ML-Trees; typical example:
1121    
1122            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1123    
1124        This currently has no further impact on SML/NJ since mlriscGen does
1125        not seem to generate such patterns in the first place, and c-calls
1126        (which did generate them in the beginning) has meanwhile been fixed
1127        so as to avoid them as well.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume
1131    Date: 2001/06/15 15:05:00 EDT
1132    Tag: blume-20010615-sparc
1133    Description:
1134    
1135    The purpose of this update is to provide an implementation of NLFFI
1136    on Sparc machines.
1137    
1138    Here are the changes in detail:
1139    
1140       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1141       the Sparc implementation of the c-calls API.
1142       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1143       for accessing the ML frame.  Thus, we have a real frame pointer and
1144       can freely modify %sp without need for an omit-frame-ptr phase.
1145       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1146       this case.
1147       * ml-nlffigen has been taught to produce code for different architectures
1148       and calling conventions.
1149       * In a way similar to what was done in the x86 case, the Sparc
1150       backend uses its own specific extension to mltree.  (For example,
1151       it needs to be able to generate UNIMP instructions which are part
1152       of the calling convention.)
1153       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1154       to make it easier to plug in new machine- and os-dependent parts).
1155    
1156    There are some other fairly unrelated bug fixes and cleanups as well:
1157    
1158       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1159       that they properly share their libraries with existing SML/NJ libraries.
1160       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1161       follow-up error messages.
1162       * Updates to CM/CMB documentation.
1163    
1164    TODO items:
1165    
1166       * MLRISC should use a different register as its asmTemp on the Sparc.
1167         (The current %o2 is a really bad choice because it is part of the
1168         calling conventions, so things might interfere in unexpected ways.)
1169    
1170    ----------------------------------------------------------------------
1171    Name: Matthias Blume
1172    Date: 2001/06/07
1173    Tag: blume-20010607-calls
1174    Description:
1175    
1176    A number of internal changes related to C calls and calling conventions:
1177    
1178    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1179       number of bytes popped implicitly (by the callee).  In most cases
1180       this field is 0 but on x86/win32 it is some non-zero value.  This
1181       is information provided for the benefit of the "omit-frameptr" pass.
1182    2. The CALL instruction on the x86 carries a similar "pops" field.
1183       The instruction selection phase copies its value from the ML-Tree
1184       CALL statement.
1185    3. On all other architectures, the instruction selection phase checks
1186       whether "pops=0" and complains if not.
1187    4. The c-calls implementation for x86 now accepts two calling conventions:
1188       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1189       up after the call and pops is set to 0.  For "stdcall", the caller
1190       does nothing, leaving the cleanup to the callee; pops is set to
1191       the number of bytes that were pushed onto the stack.
1192    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1193       distinguish between "ccall" and "stdcall".
1194    6. The UNIMP instruction has been added to the supported Sparc instruction
1195       set. (This is needed for implementing the official C calling convention
1196       on this architecture.)
1197    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1198       work with the latest CM.
1199    
1200    ----------------------------------------------------------------------
1201    Name: Matthias Blume
1202    Date: 2001/06/05 15:10:00 EDT
1203    Tag: blume-20010605-cm-index
1204    Description:
1205    
1206    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1207    
1208    1. CM can now generate "index files".  These are human-readable files
1209       that list on a per-.cm-file basis each toplevel symbol defined or
1210       imported.  The location of the index file for
1211       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1212       To enable index-file generation, set CM.Control.generate_index to true
1213       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1214    
1215       The CM manual has been updated accordingly.
1216    
1217    2. I made some slight modifications to the c-calls API in MLRISC.
1218    
1219         a) There is now a callback to support saving/restoring of
1220            dedicated but caller-save registers around the actual call
1221            instruction.
1222         b) One can optionally specify a comment-annotation for the
1223            call instruction.
1224    
1225    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1226       (For example, the comment annotation shows the C prototype of
1227        the function being called.)
1228    
1229    ----------------------------------------------------------------------
1230    Name: Matthias Blume
1231    Date: 2001/06/01 13:30:00 EDT
1232    Tag: blume-20010601-nlffi-cleanup
1233    Description:
1234    
1235    This is mostly a cleanup of MLFFI stuff:
1236    
1237       - some signature files have been put into a more exposed place
1238       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1239       - ml-nlffigen changed accordingly
1240       - tutorial updated
1241    
1242    Other changes:
1243    
1244       - author's affiliation in CM manual(s) updated
1245       - some more recognized keywords added to Allen's sml.sty
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume
1249    Date: 2001/05/25 15:30:00 EDT
1250    Tag: blume-20010525-iptr
1251    Description:
1252    
1253      - put the official 110.33-README (as it appears on the ftp server) under
1254        CVS
1255      - fixed a small bug related to incomplete pointer types in
1256        ml-nlffigen
1257      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1258        (it does not need the 'f type parameter)
1259    
1260    ----------------------------------------------------------------------
1261    Name: Matthias Blume
1262    Date: 2001/05/23 14:30:00 EDT
1263    Tag: Release_110_33
1264    Description:
1265    
1266    New version number (110.33).  New bootfiles.
1267    
1268    ----------------------------------------------------------------------
1269    Name: Matthias Blume
1270    Date: 2001/05/22 18:06:00 EDT
1271    Tag: blume-20010522-targets
1272    Description:
1273    
1274    Made install.sh use file config/targets.customized if it exists, falling
1275    back to config/targets if it doesn't.  This way one can have a customized
1276    version of the targets file without touching the "real thing", thus
1277    eliminating the constant fear of accidentally checking something bogus
1278    back into the CVS repository...  (File config/targets.customized must
1279    not be added to the repository!)
1280    
1281    ----------------------------------------------------------------------
1282    Name: Matthias Blume
1283    Date: 2001/05/22 16:30:00 EDT
1284    Tag: blume-20010522-minitut
1285    Description:
1286    
1287    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1288       struct returns.
1289    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1290       incomplete, preliminary documentation for NLFFI.
1291    
1292    ----------------------------------------------------------------------
1293    Name: Matthias Blume
1294    Date: 2001/05/14 11:30:00 EDT
1295    Tag: blume-20010514-script
1296    Description:
1297    
1298    Some bugs in install script fixed.
1299    
1300    In addition to that I also made a slight change to the NLFFI API:
1301    Functors generated by ml-nlffigen now take the dynamic library as a
1302    straight functor argument, not as a suspended one.  (The original
1303    functor code used to force the suspension right away anyway, so there
1304    was nothing gained by this complication of the interface.)
1305    
1306    ----------------------------------------------------------------------
1307    Name: Matthias Blume
1308    Date: 2001/05/11 14:35:00 EDT
1309    Tag: blume-20010511-ml-nlffi
1310    Description:
1311    
1312    I finally took the plunge and added my new FFI code to the main
1313    repository.  For x86-linux it is now ready for prime-time.
1314    
1315    There are two new subdirectories of "src":
1316    
1317      - ml-nlffi-lib:
1318           The utility library for programs using the FFI interface.
1319           Here is the implementation of $/c.cm and its associated low-level
1320           partners $/c-int.cm and $/memory.cm.
1321      - ml-nlffigen:
1322           A stand-alone program for generating ML glue code from C source
1323           code.
1324    
1325    Building ml-nlffigen requires $/ckit-lib.cm.
1326    
1327    The config/install.sh script has been updates to do the Right Thing
1328    (hopefully).
1329    
1330    Notice that the source tree for the C-Kit will not be put under "src"
1331    but directly under the installation root directory.  (This is the
1332    structure that currently exists on the CVS server when you check out
1333    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1334    
1335    Bugs: No documentation yet.
1336    
1337    ----------------------------------------------------------------------
1338    Name: Matthias Blume
1339    Date: 2001/05/09 16:35:00 EDT
1340    Tag: blume-20010509-cpscontract
1341    Description:
1342    
1343    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1344    wrapper/unwrapper elimination did not decrement usage counts and some
1345    dead variables got overlooked by the dead-up logic.)
1346    
1347    ----------------------------------------------------------------------
1348    Name: Lal George
1349    Date: 2001/05/08  17:26:09 EDT
1350    Tag: george-20010508-omit-frameptr
1351    Description:
1352    
1353    Changes to implement the omit-frame-pointer optimization to support
1354    raw C calls. For now, there is only support on the Intel x86, but
1355    other architectures will follow as more experience is gained with this.
1356    
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume
1360    Date: 2001/05/07 14:40:00 EDT
1361    Tag: blume-20010507-proxies
1362    Description:
1363    
1364    I made into "proxy libraries" all libraries that qualify for such a
1365    change.  (A qualifying library is a library that has another library or
1366    groups as its sole member and repeats that member's export list
1367    verbatim.  A proxy library avoids this repetition by omitting its export
1368    list, effectively inheriting the list that its (only) member exports.
1369    See the CM manual for more explanation.)
1370    The main effect is that explicit export lists for these libraries
1371    do not have to be kepts in sync, making maintenance a bit easier.
1372    
1373    I also added copyright notices to many .cm-files.
1374    
1375    Last but not least, I made a new set of bootfiles.
1376    
1377    ----------------------------------------------------------------------
1378    Name: Matthias Blume
1379    Date: 2001/05/04 17:00:00 EDT
1380    Tag: blume-20010504-cm-lsplit
1381    Description:
1382    
1383    0. John merged pending changes to $/smlnj-lib.cm
1384    
1385    1. Allen's previous change accidentally backed out of one of Lal's
1386       earlier changes.  I undid this mistake (re-introducing Lal's change).
1387    
1388    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1389       within the compiler where applicable.  There is some code simplification
1390       because of that.
1391    
1392    3. The "split" phase (in FLINT) is now part of the default list of phases.
1393       Compiler.Control.LambdaSplitting.* can be used to globally control the
1394       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1395       it can now also be controlled on a per-source basis: CM has been taught
1396       a new tool parameter applicable to ML source files.
1397    
1398       - To turn lambda-splitting off completely:
1399            local open Compiler.Control.LambdaSplitting in
1400                val _ = set Off
1401            end
1402       - To make "no lambda-splitting" the global default (but allow per-source
1403         overriding); this is the initial setting:
1404            local open Compiler.Control.LambdaSplitting in
1405                val _ = set (Default NONE)
1406            end
1407       - To make "lambda-splitting with aggressiveness a" the global default
1408         (and allow per-source overriding):
1409            local open Compiler.Control.LambdaSplitting in
1410                val _ = set (Default (SOME a))
1411            end
1412    
1413       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1414         write (in the respective .cm-file):
1415            a.sml (lambdasplitting:off)
1416       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1417            a.sml (lambdasplitting:on)
1418       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1419         <a> is a decimal non-negative integer):
1420            a.sml (lambdasplitting:<a>)
1421       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1422            a.sml (lambdasplitting:infinity)
1423       - To use the global default for a.sml:
1424            a.sml (lambdasplitting:default)
1425         or simply
1426            a.sml
1427    
1428    ----------------------------------------------------------------------
1429    Name: Allen Leung
1430    Date: 2001/05/04 01:57:00 EDT
1431    Tag: leunga-20010504-sync
1432    Description:
1433    
1434      MLRISC features.
1435    
1436      1. Fix to CMPXCHG instructions.
1437      2. Changed RA interface to allow annotations in callbacks.
1438      3. Added a new method to the stream interface to allow annotations updates.
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume
1442    Date: 2001/05/01 11:45:00 EDT
1443    Tag: blume-20010501-pcedittmp
1444    Description:
1445    
1446    Changed install.sh to use the current working directory instead of
1447    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1448    of /usr/tmp caused trouble with MacOS X because of file premission
1449    problems.
1450    
1451    ----------------------------------------------------------------------
1452    Name: Matthias Blume
1453    Date: 2001/04/20 11:10:00 EDT
1454    Tag: blume-20010420-inMLflag
1455    Description:
1456    
1457     - added vp_limitPtrMask to vproc-state.h
1458       (for use by the raw-C-calls mechanism to implement proper interrupt
1459        handling)
1460     - made the ML compiler aware of various data-structure offsets so it
1461       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1462     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1463       raw C-calls
1464    
1465    ----------------------------------------------------------------------
1466    Name: Lal George
1467    Date: 2001/04/20 09:15:28 EDT
1468    Tag: george-20010420-macosX
1469    Description:
1470    
1471     - Changes to port to Mac OS X; Darwin.
1472    
1473     - In the process I found that sqrt was broken on the PPC, because the
1474       fsqrt instruction is not implemented.
1475    
1476    ----------------------------------------------------------------------
1477    Name: Matthias Blume
1478    Date: 2001/04/18 12:45:00 EDT
1479    Tag: blume-20010418-ccalls
1480    Description:
1481    
1482     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1483       (this bug prevented structure arguments containing pointers from being
1484        passed correctly)
1485     - changed the raw-C-call code in mlriscGen.sml in such a way that
1486       structure arguments are represented as a pointer to the beginning
1487       of the structure (instead of having a series of synthesized arguments,
1488       one for each structure member)
1489    
1490     - made makeml script's verbosity level configurable via environment
1491       variable (MAKEML_VERBOSITY)
1492    
1493     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1494       in rawmem-x86.sml; we are now using the real thing
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume
1498    Date: 2001/03/22 16:25:00 EST
1499    Tag: blume-20010322-bootfiles
1500    Description:
1501    
1502    Created a new set of bootfiles (for your automatic installation convenience).
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2001/03/22 15:10:00 EST
1507    Tag: blume-20010322-rawmem-parcm
1508    Description:
1509    
1510    1. All "raw memory access" primitives for the new FFI are implemented now
1511       (at least on the x86).
1512    2. Some further cleanup of CM's parallel make mechanism.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume
1516    Date: 2001/03/19 17:53:00 EST
1517    Tag: blume-20010319-parallel
1518    Description:
1519    
1520    Parallel make (using compile servers) now works again.
1521    
1522    To this end, CM.stabilize and CMB.make have been modified to work in
1523    two passes when compile servers are attached:
1524       1. Compile everything, do not perform stabilization; this pass
1525          uses compile servers
1526       2. Stabilize everything; this pass does not use compile servers
1527    If there are no compile servers, the two passes are combined into one
1528    (as before).  Splitting the passes increases the inherent parallelism
1529    in the dependency graph because the entire graph including all
1530    libraries is available at the same time.  This, in turn, improves
1531    server utilization.  The downside is that the master process will
1532    have to do some extra work after compilation is done (because for
1533    technical reasons it must re-read all the binfiles during stabilization).
1534    
1535    ----------------------------------------------------------------------
1536  Name: Matthias Blume  Name: Matthias Blume
1537  Date: 2001/03/16 12:22:00 EST  Date: 2001/03/16 12:22:00 EST
1538  Tag: blume-20010316-bootfiles  Tag: blume-20010316-bootfiles
# Line 2003  Line 3523 
3523       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3524       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3525       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
3526       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3527       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3528       Dave's hands. :)       Dave's hands. :)
3529    

Legend:
Removed from v.800  
changed lines
  Added in v.1038

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0